feat: enable PyTorch custom ops by default

[njzjz-bot]

Summary

• change the DP_ENABLE_PYTORCH build-time default from 0 to 1 in the build backend
• update the install-from-source docs to describe PyTorch customized C++ OPs as enabled by default
• adjust related model/compression docs so they no longer tell users to set DP_ENABLE_PYTORCH=1 explicitly

Tests

• python3 -m py_compile backend/find_pytorch.py backend/read_env.py deepmd/pt/cxx_op.py
• git diff --check
• uv run --no-project --with dependency-groups --with packaging python smoke checks for default -DENABLE_PYTORCH=ON and explicit DP_ENABLE_PYTORCH=0 -> -DENABLE_PYTORCH=OFF

Authored by OpenClaw (model: custom-chat-jinzhezeng-group/gpt-5.5)

Summary by CodeRabbit

• New Features

  • PyTorch backend customized C++ OPs now enabled by default when building from source.

• Bug Fixes

  • Corrected error messages and installation guidance for PyTorch backend issues.

• Documentation

  • Updated installation and model documentation to reflect PyTorch backend enabled by default.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 4bc3bb60-d8fc-4e4f-8729-3465763315ac

📥 Commits

Reviewing files that changed from the base of the PR and between 95fa690 and 5ef0601.

📒 Files selected for processing (7)

• backend/find_pytorch.py
• backend/read_env.py
• deepmd/pt/cxx_op.py
• doc/freeze/compress.md
• doc/install/install-from-source.md
• doc/model/dpa2.md
• doc/model/dpa3.md

---

📝 Walkthrough

Walkthrough

DP_ENABLE_PYTORCH default changed from "0" to "1" in find_pytorch.py and read_env.py, making PyTorch C++ OP discovery enabled by default during source builds. A runtime error message and four documentation files are updated accordingly to remove the explicit DP_ENABLE_PYTORCH=1 requirement.

Changes

PyTorch Enabled by Default

Layer / File(s)	Summary
Build-time default change backend/find_pytorch.py, backend/read_env.py	DP_ENABLE_PYTORCH default changed from "0" to "1" in both find_pytorch() (docstring fix included) and get_argument_from_env, enabling PyTorch discovery and CMake argument injection when the variable is unset.
Runtime error message deepmd/pt/cxx_op.py	load_library's OSError rebuild guidance removes the DP_ENABLE_PYTORCH=1 prefix and corrects the pip command to pip install deepmd-kit --no-binary deepmd-kit.
Documentation updates doc/install/install-from-source.md, doc/freeze/compress.md, doc/model/dpa2.md, doc/model/dpa3.md	All four docs replace instructions to set DP_ENABLE_PYTORCH=1 with a statement that the customized OP library is installed by default when building from source; the documented default value for the variable is updated from 0 to 1.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• deepmodeling/deepmd-kit#5263: Adds and clarifies DP_ENABLE_PYTORCH guidance in doc/model/dpa2.md, directly overlapping with the installation instructions modified in this PR.

Suggested labels

build, Docs

Suggested reviewers

• wanghan-iapcm

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The PR title clearly and accurately summarizes the main change: enabling PyTorch custom operators by default. It directly reflects the primary objective of changing the DP_ENABLE_PYTORCH default from 0 to 1 across build system, code, and documentation.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 82.14%. Comparing base (95fa690) to head (5ef0601).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5578      +/-   ##
==========================================
- Coverage   82.15%   82.14%   -0.01%     
==========================================
  Files         900      900              
  Lines      104140   104141       +1     
  Branches     4470     4470              
==========================================
- Hits        85553    85546       -7     
- Misses      17181    17186       +5     
- Partials     1406     1409       +3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.